Regarding spawning new workers, the README says:

The app server will automatically spawn a new worker if one dies while handling a request.

This should only happen if the app was properly initialized and the workers were all forked successfully at some point in the past as it could enter a recursive loop. Moreover, the process should exit if it fails on first try.